🔥 Free Monthly Webinars (30 minutes): Check out the schedule! Extremely Popular (Don't miss out) 🔥Learn More

Warning System Speech Intelligibility Measurement

Speech intelligibility refers to the clarity and comprehensibility of spoken words to a listener. Measuring speech intelligibility is a requirement of Australian Standards AS1670.1 and AS1670.4 to establish criteria for measuring and recording the baseline data at each place specified within an evacuation zone and within a 6m radius of each required exit.

Both Australian Standard AS 1670.1 and AS 1670.4 establish the requirement to measure and ensure speech intelligibility of a building occupant warning system or emergency warning system shall be not less than 0.7 on the Common Intelligibility Scale (CIS) within the emergency zone.

At all places within the emergency zone where live speech is provided and the ambient SPL (noise) is less than 85 dB(A) as determined using the method in Appendix H, and the reverberation time (T60) is less than 1.5 s speech intelligibility shall be not less than 0.7 on the Common Intelligibility Scale (CIS).

Clause 3.22.5 of AS 1670.1 and Clause 4.9 of AS 1670.4 when discussing intelligibility also specify certain criteria where speech intelligibility need not be measured as follows;

  • The live speech signal-to-noise ratio exceeds 10 dB (per Clause H.4, Appendix H).
  • The reverberation time (T60) is less than 1.5 seconds, as shown by an acoustic report to AS/NZS 2107.
  • Loudspeakers are spaced no more than twice their mounting height from the floor.
  • Visual alarm devices (VADs) are installed as per Clause 4.6.

Speech Intelligibility Measurement Methods

Standards like AS 1670.1 and AS 1670.4 outline SIX methods (below) for measuring speech intelligibility, each with specific strengths and limitations. It is recommended to choose a method that best suits the intelligibility range being tested, such as STI for high intelligibility or phonetically balanced word scores for low levels.

Testing should follow standard procedures and reflect typical operating conditions to ensure accurate results, with corrections applied for any variations in ambient noise.

MethodDescriptionStrengthsLimitations
Speech Transmission Index (STI)Derived from modulation transfer function (MTF) measurements using computer-based systems.Provides detailed and objective measurements; widely applicable to different environments.Requires sophisticated equipment; not suitable for environments with significant non-linear distortions.
Rapid Acoustics Speech Transmission Index (RASTI)Simplified STI method using two octave-band noise carriers and limited modulation frequencies.Faster and simpler alternative to STI; useful for quick assessments.Less accurate than STI; overestimates intelligibility in some cases.
Phonetically Balanced (PB) Word ScoresUses preselected words embedded in carrier phrases, evaluated by a panel of listeners.Highly accurate for subjective testing; reflects realistic acoustic reflections and reverberations.Time-consuming and requires trained listeners; limited to subjective evaluation.
Modified Rhyme Test (MRT)Involves a panel of listeners testing intelligibility through word-based assessments.Effective for testing intelligibility using specific word-based criteria.Dependent on listener panel; may not reflect all real-world scenarios.
Articulation Index (AI)Calculates intelligibility based on sound pressure levels, accounting for masking, distortions, and reverberation.Accounts for multiple acoustic factors; suitable for comprehensive system evaluations.Does not account well for non-linear distortions or fluctuating noise.
Articulated Loss of Consonants (%ALcons)Measures percentage loss of consonants using simple words, excluding carrier phrases.Focuses on consonant intelligibility; useful in specific speech transmission scenarios.Prone to inaccuracies in reverberant or peak-clipped conditions due to exclusion of vowels and carrier phrases.

Accurate results from these methods require strict adherence to the relevant standards and careful execution of the measurement procedure. Additionally, ambient noise levels during testing must closely match normal operating conditions, or appropriate corrections must be applied to the test data.

Measuring Rapid Acoustics Speech Transmission Index (RASTI)

Measuring the Rapid Acoustics Speech Transmission Index (RASTI) of an emergency zone or space is relatively straight forward provided users have the correct methodology and equipment.

All results form part of the baseline data for the building and shall be recorded in the commissioning documentation. For subsequent tests in accordance with Australian Standard AS 1851 (every five-years), the results shall be recorded and compared with the baseline data.  Where the baseline data does not exist, the result may be used as the basis for future comparison.

StepProcedure / ObjectiveResult
1Record the background noise and Sound Pressure Level (SPL) with a signal analyser when the emergency zone is operating in its normal (typical) state or usage over a 60-second period. 
2With the location empty, measure the Speech Transmission Index for Public Address Systems (STIPA). To measure STIPA, it is necessary to use a meter and test signal that meets the requirements of the standard for sound level meters in accordance with BS EN 61672-1 with the STIPA functionality. 
3aFor new warning systems, the result shall be measured and recorded as part of the baseline data for the emergency warning system. The results shall then be converted to the Common Intelligibility Scale according to the curves set out in Figure I.1 of Australian Standard AS 1670.4.CIS > 0.7
3bFor an existing warning system, the result shall be measured, recorded, and compared to the baseline data to ensure the result is not adversely affected by any change in the emergency zone or space. Note: Minor variations between the baseline data at the time of commissioning and subsequent assessments do not indicate a failure of the test, provided the measured result is equal to or greater than 0.7 on the Common Intelligibility Scale.CIS > 0.7

Location & Number of Measurements

Measurements must be taken at a sufficient number of representative points in each coverage area, as outlined in the system specification, or at locations previously defined as part of the baseline data for each warning zone and within a 6 m radius of the approach to a required exit.

The average intelligibility value (IAV) on the CIS scale and the standard deviation (σ) must be calculated.

The value of IAV - σ must exceed the CIS limit specified of not less than 0.7 on the Common Intelligibility Scale (CIS).

If results fall within ±σ (one standard deviation) of the limit, measurements should be repeated with more sample points. The mean intelligibility and its 95% confidence interval across the area must also be determined.

    Other people visiting Firewize also found these similar articles interesting...

    Disclaimer

    Firewize policy is one of continuous improvement, details of products, procedures or methods described within this document are subject to change without notice. All information provided here is believed to be correct at the time of publication.

    Every effort has been made to ensure the accuracy of information which is provided in good faith nothing contained herein is intended to incorporate any representation or warranty, either express or implied or to form the basis of any legal relations between the parties hereto, additional to or in lieu of such as may be applicable to a contract of sale or purchase.

    This website is provided solely on the basis that users will be responsible for making their own assessment of the matters discussed herein and are advised to verify all relevant representations, statements and information. Firewize does not accept any liability for any injury, loss or damage incurred by use of or reliance on the information.